Re: Requirement 5

(cross posting, as this particular thread did start on public-xml-id)

Norm Walsh writes:

>> I further suspect that a fair number of them
>> will not have remembered the fact that 
>> unqualified names beginning with "[xX][mM][lL]" 
>> are reserved and have probably already constructed
>> elements and attributes with those names (even, 
>> conceivably, xmlid).

Just curious:  am I right in guessing that the decision to reserve the 
unqualified names may in part be a reflection of the fact that there were 
no qualified names at the time the XML Rec was published?  There is the 
statement in the original recommendation [1] that:

"Note: The colon character within XML names is reserved for experimentation 
with name spaces. Its meaning is expected to be standardized at some 
future point, at which point those documents using the colon for 
experimental purposes may need to be updated. (There is no guarantee that 
any name-space mechanism adopted for XML will in fact use the colon as a 
name-space delimiter.) In practice, this means that authors should not use 
the colon in XML names except as part of name-space experiments, but that 
XML processors should accept the colon as a name character."

This stands in somewhat interesting juxtoposition with the decision to use 
the attribute name "xml:lang" in that same edition of XML 1.0. Nonetheless 
I wonder whether the authors of XML 1.0 would have reserved anything other 
than the prefixes had a stable definition of Namespaces in XML been then 
available.  Indeed, one might argue that it would be a reasonably 
compatible change for XML 1.0 Fourth Edition and/or XML 1.1 Second Edition 
to unreserve the unqualified names.

I agree with Norm that users have developed an expectation that the 
definitions of unqualified names are scoped to the element types in which 
they appear.  While there is debate as to whether qualified attributes are 
ever so-scoped (any fans of the dreaded XML schema issue 208?), I think 
that in the particular case of the xml: prefix, it's understood that the 
definitions are invaribly global to all uses across XML documents and 
vocabularies. 

Noah

[1] http://www.w3.org/TR/1998/REC-xml-19980210

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------








Norman Walsh <Norman.Walsh@Sun.COM>
Sent by: public-xml-id-request@w3.org
02/08/2005 12:56 PM

 
        To:     Elliotte Harold <elharo@metalab.unc.edu>
        cc:     public-xml-id@w3.org, (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        Re: Requirement 5


/ Elliotte Harold <elharo@metalab.unc.edu> was heard to say:
| Placing the id attribute in the XML
| namespace, has the effect of
| inadvertently applying IDness to
| constructs which are intended not to
| have it through the intermediary of
| canonicalization. I therefore believe
| that the current draft of xml:id does
| not meet its stated requirements, and
| request that the working group revise
| the draft so that this requirement is
| satisfied before issuing it as a
| recommendation.

I find this line of argument baffling.

There are clearly bugs in the C14N spec. I don't think there's any
blame or fault associated with that, bugs happen.

  1. C14N predicted the semantics of attributes that were not
     under its control and those predictions turned out to be
     incorrect.

  2. It handles xml:base incorrectly.

The fact that xml:id causes this C14N bug to become evident in a
slightly different way than xml:base does is being used as an argument
for renaming xml:id to xmlid for the expedient of working around the
bug. Are we to fix the bug with respect to xml:base by rescinding that
specification and reissuing it using the attribute name 'xmlbase'? (I
note, though I recognize that it hardly constitutes a reasonable
argument in this case, that had the XML Recommendation named xml:lang
and xml:space, xmllang and xmlspace, respectively, that we'd probably
be in the same situation with respect to xmlid. Would the argument
then be that we should name it xml:id?)

It seems to me that C14N needs to be fixed. Fixing it will remove any
problems associated with xml:id (that would not equally be associated
with xmlid). I don't see any motivation here for renaming xml:id.

In addition, I remain of the opinion that users who will immediately
understand what xml:id is and why and how it works will be confused
and disturbed by xmlid. I further suspect that a fair number of them
will not have remembered the fact that unqualified names beginning
with "[xX][mM][lL]" are reserved and have probably already constructed
elements and attributes with those names (even, conceivably, xmlid).
(Though again, that does not constitute a reasonable argument here.)

                                        Be seeing you,
                                          norm

-- 
Norman.Walsh@Sun.COM / XML Standards Architect / Sun Microsystems, Inc.
NOTICE: This email message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution is prohibited.
If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.